This video looks at the very different incentives and concerns behind those positions. The motives of AI CEOs can reasonably be questioned — but calls for stronger oversight are not coming from them alone. Prominent AI researchers and former insiders including Geoffrey Hinton, Yoshua Bengio and Ilya Sutskever have also raised serious concerns.
That doesn’t mean every call for regulation has the same motivation. AI companies face competitive pressure, shareholder obligations and the risk that unilateral safety measures simply hand an advantage to competitors willing to take greater risks.
The deeper question may therefore be whether AI can safely remain primarily a private competitive asset at all — or whether increasingly powerful AI ultimately needs some form of international framework and a mission extending beyond shareholder profit.
AI pause? AI regulation? Self-regulation? Or a fundamentally different model for governing AI?
#AI #AIRegulation #AISafety #ArtificialIntelligence #AIGovernance
AI regulation, AI pause, AI safety, AI governance, artificial intelligence, AI self regulation, Sam Altman, Dario Amodei, Elon Musk, Donald Trump, Jensen Huang, Geoffrey Hinton, Yoshua Bengio, Ilya Sutskever, AI CEOs, AI risks, AI oversight, AI policy, AI regulation debate, artificial intelligence regulation, AI safety debate, AI industry, AI companies, AI competition, AI existential risk, AI alignment, future of AI, global AI governance, regulating AI, should AI be regulated, pause AI, AI development, AI race, AI control
With President Trump and Nvidia CEO Jensen Huang both saying AI can self-regulate, it’s easy to dismiss calls by AI CEOs Altman, Amodei and Musk as unrealistic ploy to drive up, investor interest in the power of AI.
But then you have those who quit, claiming to fear the risks. What’s in it for them?
Jacob Coxon, who kicked off the recent debate, was just mid-level and new to Anthropic.
But he’d been at OpenAI three years and seemingly moved in search of better oversight but didn’t find it.
But his calls have been supported by far more senior people.
Some still there, but others like Joe Benton, former senior manager of an oversight team, said they also left over safety concerns.
Godfather of AI, Geoffrey Hinton, left his position at Google in 2023 to speak freely about the potential risks of AI.
Since then, he’s shared several key perspectives on AI safety.
There’s also Yoshua Bengio, considered another godfather of AI.
Ilya Sutskever, former chief scientist at OpenAI, I could go on.
Neither I, nor anyone listening can really feel that they’re better to judge the risks, than this list. And they’re concerned. Those CEOs, they’re not all the same and they may have different mixes of reasons.
But without outside regulation, they have a fiduciary duty to maximise value for their shareholders.
Plus, self-regulation is pointless, if a competitor without the same safeguards could both gain dominance and expose the world to the dangers you’re taking steps to prevent.
Their calls for regulation may be pointless, and without trusted international cooperation, go nowhere. Maybe all they want is to try and keep those altruistic employees onside.
But this is clearly a serious issue.
My own view, and I’m far from alone, and it includes at least several of those mentioned as AI founders, is that AI needs to be a global public asset and needs to have a clear mission other than shareholder profit.
We widely view capitalism as the best system there is, but also with good reason, not one country runs everything under capitalism or for profit.
Capitalism may be best as a system, but it’s not best for everything!
Consider:
From on screen in video
- Public Libraries.
- Fire Brigades.
- Police.
- The Justice system.
- …
- Public Land!
If you were to calculate the total value of all public land ( including that not normally counted, like sidewalks/footpaths ), it would dwarf the global total stock market value.
Simple example, we don’t pay fees to companies owning all the footpaths and sidewalks, but instead have a concept of public land.
AI may end up needing to be in that same category. Something to consider. Interesting to hear your thoughts.
